Role of preoperative PROMIS scores in predicting postoperative outcomes and likelihood of achieving MCID following reverse shoulder arthroplasty

Abstract

BackgroundThe patient-reported outcomes measurement information system (PROMIS) has emerged as an efficient and valid outcome measure in various shoulder surgeries. The purpose of this study was to investigate the influence of preoperative PROMIS scores in predicting postoperative PROMIS scores and the likelihood of achieving a minimal clinically important difference (MCID) following primary reverse total shoulder arthroplasty for cuff tear arthropathy. We hypothesize that preoperative PROMIS scores will influence both postoperative PROMIS scores and the probability of achieving MCID. Methods73 patients undergoing reverse shoulder arthroplasty by a board-certified shoulder and elbow surgeon were given three PROMIS CAT forms: PROMIS Upper Extremity Physical Function CAT v2.0 (“PROMIS-UE”), PROMIS Pain Interference v1.1 (“PROMIS-PI”), and PROMIS Depression v1.0 (“PROMIS-D”).). PROMIS CAT domain t scores were assessed for significance between both time points using a Paired Samples t test. Minimal clinically important difference (MCID) was calculated using the distribution method and each PROMIS domain was subsequently assessed for its discriminatory ability in predicting postoperative improvement equal to or greater than the MCID through receiver operating characteristic (ROC) curve analysis. ResultsOur cohort consisted of 73 patients (49.3% male) and an average age of 69.7 years (standard deviation, 11.9). Mean follow-up time point was 9.6 months (standard deviation, 5.0) after surgery. Preoperative PROMIS-UE, PROMIS-PI, and PROMIS-D were 29.5 ± 6.2, 63.3 ± 5.4, and 50.1 ± 9.2, respectively. Each domain significantly improved at 10-months, on average, to 40.9 ± 7.8, 51.4 ± 8.5, 42.6 ± 8.1, respectively. Following the distribution-based method for MCID calculation, we found the following MCID values for PROMIS-UE, PROMIS-PI, and PROMIS-D: 3.1, 2.7, and 4.6, respectively. ROC analysis revealed strong predictive ability for PROMIS-UE (AUC = 0.717, p < 0.05), moderative predictive ability for PROMIS-PI (AUC = 0.634, p < 0.05), and excellent predictive ability for PROMIS-D (AUC = 0.864, p < 0.05). Specifically, preoperative cutoff values of <26.0, >70.0, and >52.5 for PROMIS-UE, PROMIS-PI, and PROMIS-D are especially predictive of achieving MCID. ConclusionsPreoperative baseline scores can serve as strong predictors of success in patients undergoing primary reverse shoulder arthroplasty and can be used to both counsel patients on surgery and to tailor postoperative protocols. Level of evidenceLevel II.